Lecture Notes in Mathematics (cid:49)(cid:56)(cid:57)(cid:56) Editors: J.-M.Morel,Cachan F.Takens,Groningen B.Teissier,Paris Horst Reinhard Beyer Beyond Partial Differential Equations On Linear and Quasi-Linear Abstract Hyperbolic Evolution Equations (cid:65)(cid:66)(cid:67) Author HorstReinhardBeyer CenterforComputation&Technology LouisianaStateUniversity (cid:51)(cid:51)(cid:48)JohnstonHall BatonRouge LA70803 USA e-mail:[email protected] LibraryofCongressControlNumber:(cid:50)(cid:48)(cid:48)(cid:55)(cid:57)(cid:50)(cid:49)(cid:54)(cid:57)(cid:48) MathematicsSubjectClassification((cid:50)(cid:48)(cid:48)(cid:48)(cid:41)(cid:58)(cid:52)(cid:55)(cid:74)(cid:51)(cid:53)(cid:44)(cid:52)(cid:55)(cid:68)(cid:48)(cid:54)(cid:44)(cid:51)(cid:53)(cid:76)(cid:54)(cid:48)(cid:44)(cid:51)(cid:53)(cid:76)(cid:52)(cid:53)(cid:44)(cid:51)(cid:53)(cid:76)(cid:49)(cid:53) ISSNprintedition:(cid:48)(cid:48)(cid:55)(cid:53)(cid:45)(cid:56)(cid:52)(cid:51)(cid:52) ISSNelectronicedition:(cid:49)(cid:54)(cid:49)(cid:55)(cid:45)(cid:57)(cid:54)(cid:57)(cid:50) ISBN-10 (cid:51)(cid:45)(cid:53)(cid:52)(cid:48)(cid:45)(cid:55)(cid:49)(cid:49)(cid:50)(cid:56)(cid:45)(cid:55)SpringerBerlinHeidelbergNewYork ISBN-13 (cid:57)(cid:55)(cid:56)(cid:45)(cid:51)(cid:45)(cid:53)(cid:52)(cid:48)(cid:45)(cid:55)(cid:49)(cid:49)(cid:50)(cid:56)(cid:45)(cid:53)SpringerBerlinHeidelbergNewYork DOI(cid:49)(cid:48)(cid:46)(cid:49)(cid:48)(cid:48)(cid:55)/(cid:57)(cid:55)(cid:56)(cid:45)(cid:51)(cid:45)(cid:53)(cid:52)(cid:48)(cid:45)(cid:55)(cid:49)(cid:49)(cid:50)(cid:57)(cid:45)(cid:50) Thisworkissubjecttocopyright.Allrightsarereserved,whetherthewholeorpartofthematerialis concerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation,broadcasting, reproductiononmicrofilmorinanyotherway,andstorageindatabanks.Duplicationofthispublication orpartsthereofispermittedonlyundertheprovisionsoftheGermanCopyrightLawofSeptember(cid:57), (cid:49)(cid:57)(cid:54)(cid:53),initscurrentversion,andpermissionforusemustalwaysbeobtainedfromSpringer.Violationsare liableforprosecutionundertheGermanCopyrightLaw. SpringerisapartofSpringerScience+BusinessMedia springer.com (cid:176)c Springer-VerlagBerlinHeidelberg(cid:50)(cid:48)(cid:48)(cid:55) Theuseofgeneraldescriptivenames,registerednames,trademarks,etc.inthispublicationdoesnotimply, evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevantprotectivelaws andregulationsandthereforefreeforgeneraluse. TypesettingbytheauthorandSPiusingaSpringerLATEXmacropackage Coverdesign:WMXDesignGmbH,Heidelberg Printedonacid-freepaper SPIN:(cid:49)(cid:49)(cid:57)(cid:54)(cid:50)(cid:49)(cid:55)(cid:53) VA(cid:52)(cid:49)(cid:47)(cid:51)(cid:49)(cid:48)(cid:48)/SPi (cid:53)(cid:52)(cid:51)(cid:50)(cid:49)(cid:48) DedicatedtoGodtheFather Preface SemigroupTheoryusesabstractmethodsofOperatorTheorytotreatinitialbound- aryvalueproblemsforlinearandnonlinearequationsthatdescribetheevolutionofa system.Duetothegeneralityofitsmethods,theclassofsystemsthatcanbetreated inthiswayexceedsbyfarthosedescribedbyequationscontainingonlylocaloper- atorsinducedbypartialderivatives,i.e.,PDEs.Inparticular,thatclassincludesthe systemsofQuantumTheory. Another important application of semigroup methods is in field quantization. Simple examples are given by the cases of free fields inMinkowski spacetime like Klein-Gordon fields, the Dirac field and the Maxwell field, whose field equations aregivenbysystemsoflinearPDEs.Thesecondquantizationofsuchafieldreplaces the field equation by a Schro¨dinger equation whose Hamilton operator is given by thesecondquantizationofanon-localfunctionofaself-adjointlinearoperator.That operator generates the semigroup given by the time-development of the solutions of the field equation corresponding to arbitrary initial data as a function of time. Moregenerally,inthesecasesthestructuresusedintheformulationofawell-posed abstract initial value problem for the field equation also provide the mathematical framework for the quantization of the field. Quantum Theory is an abstract theory, thereforeitshouldbeexpectedthatonlyanabstractapproachtoclassicalfieldequa- tions using methods from Operator Theory is capable of providing the appropriate structuresforquantizationinthelesssimplecasesofnonlinearfields,likethegravi- tationalfielddescribedbyEinstein’sfieldequations. Ademonstrationofthestrengthofsemigroupmethodscanbeseeninthefirstrig- orousproofofwell-posedness(localintime)oftheinitialvalueproblemforquasi- linear symmetric hyperbolic systems by T. Kato in 1975 in [110]. This result is a particular application of a theorem on the well-posedness of the initial value prob- lemforabstractquasi-linearequations1,whichhasbeensuccessfullyappliedalsoto Einstein’s equation’s [102], the Navier-Stokes equations, the equations of Magne- tohydrodynamics [109] and more. To my knowledge, there is no other approach to quasi-linearequationsleadingtoatheoremofsuchgenerality. 1SeeTheorem11.0.7below. VIII Preface The semigroup approach goes beyond that of a tool for deciding the well- posedness of initial boundary value problems. For a autonomous nonlinear equa- tion the important question of the linearized stability of a particular solution leads onaspectralproblemfortheoperatorgeneratingthesemigroupgivenbythetime- development of the solutions of the linearized equation around that solution corre- spondingtoarbitraryinitialdataasafunctionoftime.2Thesemethodsalsoprovide, forautonomouslinearequations,arepresentationofthesolutionoftheinitialvalue problemasanintegralovertheresolventoftheinfinitesimalgeneratoroftheasso- ciated semigroup.3 The resolvent operators can often be represented in the form of integraloperatorswithkernelswhicharedefinedintermsofspecialfunctions.This is not only true in simple cases where the generator is a partial differential opera- torwithconstantcoefficients,butalsoinanumberofcasesinvolvingnon-constant coefficients.Inthisway,anintegralrepresentationofthesolutionoftheinitialvalue problemisachievedforallsuchcases.4Finally,semigroupmethodsprovideaframe- work which is general enough to include numerical forms of evolution equations, openingthepossibilityofcomputingtrueerrorestimatestotheexactsolution,rather thanresidualerrors. Thisshouldnotgivetheimpressionthatthesemigroupapproachcouldreplaceall ‘hard’analysisfacts.Instead,itreducessuchapplicationtoabareminimum,which givestheapproachitsefficiency.5Forinstance,resultsfromharmonicanalysisorthe theoryofsingularintegraloperatorshaveapplicationsinsocalled‘commutatores- timates’wherethecommutatorofanintertwiningoperatorwiththeprincipalpartof apartialdifferentialoperator,usuallytwounboundedoperators,hastobeestimated. Toachievemostgeneralresults,itisoftennecessarytochooseintertwiningoperators asnon-localoperators. These methods are especially attractive if not inevitable for theoretical physi- cists in view of their comprehension of classical physics and quantum physics. In addition, their efficiency is of advantage in view of time restrictions in the math- ematics education of physicists. In spite of their power, efficiency and versatility, semigroupmethodsaresurprisinglylittleusedintheoreticalphysics.6,7Thisappears toberelatedtotwomisconceptions. First, because of the requirements of Special Relativity and General Relativity, currentproblemsinfundamentaltheoreticalphysicsnecessarilyleadonhyperbolic 2Forinstance,seeChapter5.4. 3Roughlyspeaking,foraprecisestatementsee,e.g.,Chapter4.3. 4See,e.g.,[25]. 5Butitisthebeliefoftheauthorthatthenecessityoftheuseof‘hard’toachieveanalyt- ical facts in the solution of a problem indicates that its structure has not yet been fully understood. 6Paradoxically, in some sense, one could also hold the opposite view that these methods havebeenusedforalongtimeintheoreticalphysics,mostlywithoutrealizingthattheyare rootedininSpectralTheory. 7Inaddition,theauthorisnotawareofasingleintroductiontoPDEbasedsolelyonSemi- groups/OperatorTheory,althoughthiswouldhavebeenpossibleevenbeforetheappear- anceofclassicalintroductionslike[20]thatuselessgeneralmethods. Preface IX partialdifferentialequationsystems(‘hyperbolicproblems’).Standardtextsonsemi- groupsoflinearoperatorsmainlyfocusonapplicationstoparabolicpartialdifferen- tialequationsystems(‘parabolicproblems’).Differentlytothehyperboliccase,this leads to the consideration of strongly continuous semigroups which are in addition analytic. Apparently, the consideration of parabolic problems originates from a fo- cusonengineeringapplications.Engineeringsciencespredominantlyapplyclassical Newtonian physics where signal propagation speeds are not limited by the speed of light in vacuum as this is the case in Special Relativity/General Relativity. For example,theevolutionofacompactlysupportedtemperaturefieldinspaceattime t “ 0 by the parabolic heat equation leads to a temperature field that has no com- pact support for every t ą 0. As consequence, in the evolution signal propagation speeds occur that exceed the speed of light in vacuum, and hence the equation is incompatible with Special Relativity. The same is also true for Schro¨dinger equa- tions.8 The evolution of hyperbolic partial differential equations systems preserves thecompactnessofthesupportofthedataundertimeevolution.Tomyexperience, the focus on engineering applications has lead to the quite common misconception among physicists, and to some extent also among mathematicians working in the field of partial differential equations, that semigroup methods cannot be applied to hyperbolicproblems. Second,tomyexperience,anothercommonmisconceptionisthatsemigroupsof linearoperatorscanonlybeappliedtosystemsoflinearpartialdifferentialequations. Thismightbeinfluencedbythefactthatmoststandardtextsonsemigroupsoflinear operators,indeed,focusmainlyonsuchapplications. Asaconsequence,thecourseshouldleadasrapidlyaspossiblefromautonomous linearequationstothenonlinear(quasi-linear)equationswhicharenowseeninthe hyperbolic problems emanating from current physics. Particular stress is on wave equationsandHermitianhyperbolicsystems.Thelastcovertheequationsdescribing interacting fields in physics and therefore the major part of nonlinear equations occurring in fundamental physics. Throughout the course applications to prob- lemsfromcurrentrelativistic(‘hyperbolic’)physicsareprovided,whichdisplaythe potentialofthemethodsinthesolutionofcurrentproblemsinphysics.Theseinclude problems from black hole physics, the formulation of outgoing boundary condi- tionsforwaveequationsandthetreatmentofadditionalconstraints.Thelasttwoare important current problems in the numerical evolution of Einstein’s field equations for the gravitational field. Some of the examples contain new unpublished results oftheauthor.Tomyknowledge,themajorpartofthematerialinthesecondpartof thenotes,includingnon-autonomousandquasi-linearHermitianhyperbolicsystems, hasappearedonlyinsideresearchpapers. The orientation of this course towards abstract quasi-linear evolution equations made it necessary to omit a number of topics that were not directly important for achieving itsgoals. Onthe other hand, texts on semigroups of linear operators that considerthosetopicsareavailable[39,47,52,57,90,99,106,120,168,179,224].Also, 8ThiswasthereasonforthedevelopmentofQuantumFieldTheory. X Preface thereisatheoryofnonlinearsemigroupsthatlargelyparallelsthatofsemigroupsof linearoperators[15,19,88,138,146,167]. This course assumes some basic knowledge of Functional Analysis which can be found, for instance, in the first volume of [179]. Some examples assume more specializedknowledgeofpropertiesofself-adjointlinearoperatorsinHilbertspaces whichcanbefound,forinstance,inthesecondvolumeof[179].Inaddition,some applications assume basic knowledge of Sobolev spaces of the L2-type as is pro- vided,forinstance,in[217].Otherwisethiscourseisself-contained.Thematerialis presentedinascompressedaformaspossibleanditsresultsareformulatedinview towards applications. In general, theorems contain their full set of assumptions, so that a study of their environment is not necessary for their understanding. For this reason and to limit the size of this text, shorter definitions appear as part of theo- rems.Inaddition,theabstracttheoryanditsapplicationsappearinseparatechapters toallowthereadertoestimatethenecessarytimetoacquirethetheory. Chapters2–5constitutethenotesofacoursegivenattheDepartmentofMathe- maticsoftheLouisianaStateUniversityinBatonRougeinSpring2005.Theypro- vide a basic introduction into the properties of strongly continuous semigroups on Banach spaces and applications to autonomous linear hyperbolic systems of PDEs from General Relativity and Astrophysics. The theoretical part is kept to a mini- mumwithaviewtoapplicationsinthefieldofhyperbolicPDEs.Itisformulatedin a way which is expected to be natural to readers with a knowledge of the spectral theoryofself-adjointlinearoperatorsinHilbertspaces.Anexception,tothisrestric- tiontotheminimuminthesechapters,isthetreatmentoftheintegrationofBanach space-valued maps which is more detailed than is usual in most other comparable texts.Thisisduetothefactthatsuchintegrationisabasictoolwhichisusuallynot coveredinstandardFunctionalAnalysiscourses,atleastnottoanextentneededin thestudyofsemigroupsofoperators.Otherwise,thetheoreticalmaterialisstandard andforthisreasononlyfewreferencestoliteraturearegiven.Formorecomprehen- siveintroductionsintothetheoryofsemigroupsoflinearoperatorsthatgiveamore exhaustivelistofreferences,thereaderisreferred,forinstance,to[57,90,99,168]. Chapters6–12arethenotesofasubsequentcourseatthesameplaceinSpring 2006.Theyintroduceintothefieldofabstractevolutionequationswithapplications tonon-autonomouslinearandquasi-linearhyperbolicsystemsofPDEs.Thissecond part of the notes follows closely the late Tosio Kato’s 1993 paper ‘Abstract evolu- tionequations,linearandquasilinear,revisited’inProceedingsoftheInternational Conference in Memory of Professor Kosaku Yosida held at RIMS, Kyoto Univer- sity,Japan,July29-Aug.2,1991,[114].ThoseresultsaremoregeneralthanKato’s well-knownolderresults9 [107–109]inthattheydon’tassumespecialpropertiesof theunderlyingBanachspaces.ProofsofKato’sresultsareaddedalongwithdetailed examples displayingtheirapplication toproblems incurrent relativisticphysics. In afewplaces,Lemmatawereaddedtohisoutlinethatappearednecessaryforthose proofs. 9Thoseresultsassumethereflexivityandinsomeplacesalsotheseparabilityoftheunder- lyingBanachspaces. Preface XI Acknowledgments IamespeciallyindebtedtoOlivierSarbach,whodiligentlyreviewedthemajorpart ofthenotesandrecommendedvaluablemodificationsandcorrections.Also,Ithank FrankNeubranderandStephenShipmanforilluminatingdiscussionsinconnection with the notes. Finally, I thank all who have worked on the book, especially the editorialandproductionstaffofSpringer-Verlag. BatonRouge,September2006 HorstBeyer
Description: